PROJECT SUMMARY For several decades, researchers have studied the neural mechanisms of spatial attention (i.e., the enhancement of neural activity for specific spatial locations) and target similarity (i.e., the enhancement of neural activity for specific inputs that look like a search target) in prefrontal cortex. The vast majority of this research has been carried out with tasks using impoverished stimuli?typically a small number of targets and distractors that abruptly flash against a uniform background. This is very different from sensory stimulation in natural environments, where boundaries between objects are less clear, the background is more complex, and most objects remain relatively static. Furthermore, artificial search tasks typically require just one saccade to a rewarding target, while natural search behavior consists of many saccades, each with different strategic goals (e.g., to explore and find potential targets, or to exploit known targets). It is currently not known how attention operates in these natural conditions. Therefore, the overarching goal of this research is to understand how attention is allocated in the real world. Neural activity from the prefrontal cortex (frontal eye fields) will be recorded while monkeys perform search tasks at different levels of abstraction from natural scene search. Specifically, the goal Aim 1 is to determine how the spectrum of saccadic strategies during natural scene search (e.g., explore or exploit) influences the allocation of attention. While monkeys freely search for a target embedded in a background of patterned noise, the allocation of attention will be flexibly and continuously sampled by rapidly (5-10 Hz) flashing a salient probe at a random location. The goal of Aim 2 is to test whether a map of target-similarity is encoded in prefrontal cortex before an object is recognized and attended spatially (i.e., target similarity guides spatial selection), or whether it is dependent upon prior allocation of spatial attention toward a detected object. This will be accomplished by constraining saccadic behavior while providing naturalistic sensory stimulation. The monkeys will be trained to detect and discriminate a target that gradually emerges from a noisy background. The target will be psychophysically detectable before the monkey detects it, allowing us to test whether FEF encodes target similarity even when spatial selection has not occurred. Together, these experiments will provide key insights into the dynamics of spatial attention and target similarity when sensory stimulation is natural and saccadic search behavior is unconstrained.